Hendrew S. Girotto, Raphael Issamu Tsukada, Sávio S.V. Vianna

We propose a new approach to optimise the number and location of gas detectors employing the weighted set covering problem combined with computational fluid dynamics (CFD). The optimisation mesh that represents the nodes and the links of the Graph $V\{N,L\}$ ($N$ is the set of nodes and $L$ is the set of links) of the set covering problem was calculated using computational fluid dynamics. The 0–1 integer programming model is solved using Balas algorithm. The developed model covered 100% of the area. We also tested the model in a typical chemical process module and the same coverage was attained. The optimal distribution of gas detectors was able to detect the ammonia release within 5 s. We also show that the procedure can be combined with the leak frequency for the determination of the weights of the subareas (nodes of the Graph $V\{N,L\}$). Numerical findings show that the weighted approach led to a shorter detection time when compared with the case where the subareas were not assigned a weight.